cttet

Polymarket is banned in my country, but I would really like to bet on the other side. Just like second order can provide much information than first order, the computation cost forbids it. There could be a way that is easier to compute/parallalize that overpowers the benifit of gradient. I don't think the current methods like EA/Dust is the solution either, though, they are just imitating gradient.

In modal logic sense, Chinese is inherently more □ oriented while the language in US is using ◇ more. so in Chinese ¬(□¬A) can be used to represent a possibly concept ♢A

Yep, all users are logged out automatically...

A "worst image" instead of best image competition may be easy to implement and quite indicative of which one has less frustration experience.

I rarely think with words, when I think with words it is like 20x slower, it is more robust, but in that case I would use pen and paper for that.

At least most functional language tutorial claim to be based on abstract machines, not like the C language, which is a spherical cow that people not often aware of. https://queue.acm.org/detail.cfm?id=3212479

Thanks! Really like your intuition!

Oh it is selected output, yes I meant that I was a bit confused. So in the initial design when you first tried it, you passed both to the next layer? or it is part of where you find out to perform better?

It seem to have both feature and a discrete number passed into next layer, which one did you think of first? or it is both by design?

It may be a Claude specific thing. I tried to ask Claude to various tasks in machine learning, like implement gradient boosting without specifying the language, thinking it will use Python since it is the most common option and have utilities like Numpy to make it much easier. But Claude mostly choose Javascript for the language and somehow managed to do it in JS.

Maybe it is a cultural difference aspect, but I feel that "supermarket workers, gas station attendants" (in an Asian country) that I know of should be quite capable of most ARC tasks.

The point is not that having a high score -> AGI, their ideas are more of having a low score -> we don't have AGI yet.

In all their experiments, backprop is used for most of their parameter though...

Since a long time ago, and I found a trend that in more theoretical fields, the more 'clickbaity' the titles are, e.g. TCS, machine learning theory, etc.

Never read his books. But I found good architectures is very helpful for LLM-assisted coding, if I keep things nicely named and decoupled.

If NIH syndrome boosts morale of the team, it should be helpful on overall team progress though.

> people like me can't stop commenting on it

Yeah this loop will go on with this comment, it seems.

Paper enables non-vision based(scroll bar) random access to content, when you keep going back an forth between two pages, it is very annoying on all current devices except paper. Vision pro/VR/AR or a particular multi-screen set-up can achieve that, but so far all alternatives are not as good.

For me bookmarking does not mean that I want to read the article, it usually mean that I think it may be useful for certain task in the future want it to be in my search result some times. Since sometimes I want to read in depth about something that I read briefly long ago, it is really hard for me to find it back.

Looks great! I wonder is it API based or utilising a local LLM?

'source available' also implies the case the source is only available upon request. e.g. Microsoft provided the full source code of Office to China, it is also "source available".

I don't consider the comments helpful as well, but quite contrastively, because of the over optimism.

Lots of the comments contribute the failure to quite a few reasons, and emphasize on the effect of the reasons. But the real thing is this actually can imply that the failure can be narrow down to "a few simple obvious fixes".

To me it make me feel good since "a bad title" can be a very simple excuse to think that "what if I fix that, things would get better" and ignore more problem in depth..

Things like "It's OK, you have tried your best" is the real negativity to me, since it implies that I have reached the best I can do in potential/capacity, and can be quite discouraging. But it is quite helpful sometimes since I can recogonize my limits and try to improve myself or pivot.

So the questionis: if you don't refer to dictionaries, how to determine the meaning of the words?

A. If you mean that there is some authority that can do that, what authority other than a dictionary? Alternatively, B. if you mean that there shouldn't be any authority that can define the meaning of the word. In this way you can't stop people to understand it as it's literal meaning and blame them.

I think people just don't understand that end user API design is a user preference thing, not a universal thing... And exposing what kind of level on a hardware is just the same as API design... And some people just keep having the ideas of "I'd like API to be like A, so there wouldn't be possible any people liking another API B..."

> “Open Source” is in the name only.

Hard to imagine this can be upvoted to this high.

"Open Source" just means that the source code is open, not closed.

Some people just like the source code to be open (e.g. for security reasons to examine the source code.).

It is just as simple as that.

The model is created by one person, he does not have enough time to write the paper

Interesting that you call them "fixer"s, but the point to "fix" is that the meaning is not the same and they want to change it right? If they are the same, why would they have the incentive to "fix" it?

No, it is just that you are browsing the news that only follow the trends I think. Follow more independent thinkers and things would get better.

code != imperative languages, though. There are expressible programming languages/library that can express the same semantic as math.

I will recommend Information Theory, Inference, and Learning Algorithms by David MacKay. If you really want to understand the "learning" part, rather than being given a methodology without knowing why or in the unsorted bounds that "guarantee" abstract things that may mismatch reality.